首页科技快讯黄仁勋：全世界50%的AI人才为什么都是中国人？

黄仁勋：全世界50%的AI人才为什么都是中国人？

来源：晰数塔互联网快讯时间：2026年07月23日 07:32

（来源：财经会议圈）

编辑：米奇

英伟达 CEO 黄仁勋在 Lex Fridman 深度播客访谈中抛出广受热议的判断：

全球约半数 AI 研究者为华裔，多数扎根中国本土。

他系统拆解中国科技产业快速崛起的底层逻辑：

扎实的数理教育、城市与企业间充分竞争、工程师圈层知识互通、浓厚开源氛围，并将中国定义为 “建设者之国”。同时他理性看待中美 AI 赛道，认为竞争可以驱动创新，技术脱钩并无益处。

完整中文接续译文

莱克斯・弗里德曼：站在全球 AI 长期竞争视角，这样的人才优势，是否会让中国有机会走到通用人工智能的前沿？

黄仁勋：人才是一切的起点。你无法忽视规模如此庞大的研究群体。当然，创新不单单依靠人才，算力基础设施、软件生态、真实产业场景缺一不可。但任何一场技术竞赛，人才都是最底层的根基。试图隔绝、无视如此庞大的研究者群体，长期来看并不现实。

中美在 AI 领域的竞争不可避免，而竞争本身是健康的，竞争会倒逼双方加速创新。但完全脱钩并无必要，技术本身是中立的。最好的局面，是双方持续创新、互相借鉴，共同推动人类知识边界向前延伸。

很多人问我，未来开源还是闭源模型会占据主流。我的观点很明确：开源潮流无法阻挡。当数十万优秀研究者分布在全球不同地区，开放交流本就是知识演进的自然方式。中国开发者社区对开源的参与度非常高，这一趋势只会持续增强。

我也想澄清一个容易产生的误解。我说全球大约一半 AI 研究者是华人，指的是华裔背景研究者，其中大量人员在中国本土开展研究，但也有相当一部分分布在北美、欧洲各地。人才本就会跨国流动，限制人才流动，只会拖慢全人类的技术进步。

莱克斯・弗里德曼：你把中国称作 “建设者之国”，造就它擅长打造大规模科技产业最核心特质是什么？

黄仁勋：建设者重视落地执行，专注把想法变成实体产品、工厂、整套系统。不少中国政策制定者拥有工程师背景，思考问题习惯于落地、部署、持续迭代优化，而非停留在纯理论辩论。

很多西方社会习惯在动手建设之前，花费大量时间反复推演风险。在中国，大家倾向先落地搭建，快速迭代，在发展过程中管控风险。这种思维模式让产业能够极快扩张。叠加庞大本土市场用来验证产品，新技术可以快速完成商业验证。

没有任何一个国家能够永久保有全部优势。美国在基础创新、原创科学突破层面具备极强实力；中国擅长产业规模化、工程落地、供应链协同。倘若双方都能良性发展，整个世界都会受益。

Lex Fridman：From the perspective of long-term global AI competition, does this talent advantage give China a path toward the frontier of AGI?

Jensen Huang：Talent is where everything begins. You cannot ignore a research community of this scale. Of course, innovation requires far more than just talent. Computing infrastructure, software ecosystems, real-world industry scenarios are all indispensable. But in any technology race, talent forms the fundamental foundation. Trying to isolate or disregard such a massive group of researchers is unrealistic over the long run.

Competition between the United States and China in AI is inevitable, and competition is healthy. Competition drives both sides to innovate faster. However, decoupling is unnecessary. Technology is neutral. The best outcome is that both sides continue to innovate, learn from each other, and jointly push the frontier of human knowledge forward.

Many people ask me whether open source or closed models will dominate in the future. My view is simple: open source cannot be stopped. When you have hundreds of thousands of talented researchers spread across different regions, open exchange is the natural way knowledge evolves. China’s developer community is highly active in open source, and that dynamic will only grow stronger.

I also want to clarify one misunderstanding. When I say roughly half of global AI researchers are Chinese, this refers to researchers of Chinese origin, many are working inside China, but quite a number are also distributed across North America, Europe and elsewhere. Talent flows freely across borders. Blocking talent mobility will only slow global progress.

Lex Fridman：You call China a Builder Nation. What core trait makes it so effective at building large-scale technology industries?

Jensen Huang：Builders focus on execution. They care about turning ideas into tangible products, factories, systems. Many Chinese policymakers have engineering backgrounds. They naturally think in terms of construction, deployment, iterative improvement, rather than only theoretical debate.

In many Western societies, we spend a great deal of time debating risk before we build anything. In China, people tend to build first, iterate rapidly, adjust risks along the way. That mindset enables extremely fast scaling. Combined with a huge domestic market to test products, you get rapid commercial validation for new technology.

No country can sustain all advantages forever. The United States has incredible strengths in foundational innovation, original scientific breakthroughs. China excels at industrial scaling, engineering implementation, supply chain coordination. The world benefits when both sides can flourish.